Morphology, mechanical properties, and dimensional stability of wood

particle/high density polyethylene composites: Effect of removal

of wood cell wall composition
Rongxian Ou
a,b
, Yanjun Xie
a
, Michael P. Wolcott
b
, Shujuan Sui
a
, Qingwen Wang
a,
a
Key Laboratory of Bio-Based Material Science and Technology (Ministry of Education), Northeast Forestry University, Harbin 150040, China
b
Composite Materials and Engineering Center, Washington State University, Pullman, WA 99164, USA
a r t i c l e i n f o
Article history:
Received 24 October 2013
Accepted 9 February 2014
Available online 18 February 2014
Keywords:
Wood plastic composites
Cell wall composition
Mechanical properties
Water resistance
Interfacial adhesion
a b s t r a c t
Variation in the chemical composition of wood cell walls has a signicant inuence on the properties of
wood plastic composites (WPCs). This study investigated the effect of removal of hemicellulose and/or
lignin on the mechanical properties and dimensional stability of WPCs. Four types of wood particles with
various compositions including native wood our (WF), hemicellulose-removed particle (HR), holocellu-
lose (HC), and a-cellulose (aC) were prepared and compounded with high density polyethylene (HDPE) in
an extruder, both with and without maleated polyethylene. Injection molding was used to make test
specimens. The HR-based composites exhibited the best water resistance. The HC-based composites
obtained a greater tensile modulus but a lower water resistance. The highest values for tensile strength,
elongation at brake, toughness, and impact strength were achieved by the composites lled with aC.
2014 Elsevier Ltd. All rights reserved.
1. Introduction
Lignocellulosic bers exhibit many advantages over synthetic
bers (e.g. carbon, glass and aramid), such as biodegradability,
renewability, wide availability, low density, low cost, and high spe-
cic strength and stiffness [1,2]. Wood plastic composites (WPCs)
made from lignocellulosics and thermoplastic polymers have
emerged as viable bio-based composites with low water absorp-
tion, recyclability, high yield efciency, and various geometries in
a cross section [3,4]. The ultimate performance of WPCs depends
on the intrinsic properties of lignocellulosic llers such as stiffness
of structure and the chemical composition (cellulose, hemicellu-
lose, and lignin) of cell walls.
Lignin is a branched hydrophobic heteropolymer that provides
mechanical strength to plant cell walls. For composite applications,
lignin can serve as a nucleating agent [5], antioxidant [6,7], com-
patibilizer [8,9] or even base matrix [10,11]. Karimi et al. [12]
investigated hot-pressed polypropylene (PP) composites lled with
delignied hornbeam ber by a Kraft pulping process at 170 C.
They showed that delignication improved the tensile strength,
tensile modulus, and water resistance of the resulting composites.
Fabiyi et al. produced high density polyethylene (HDPE) [13] and
polyvinyl chloride [14] composites from pine, extractives-free
pine, and pine holocellulose (delignied). They found that holocel-
lulose-based composites showed the lowest change in total color
and lightness for both outside and xenon-arc weathering. Beg
and Pickering [15] found that Kraft wood ber (Pinus Radiata) with
higher residual lignin content enabled creation of PP composites
with lower tensile and impact strength.
Hemicelluloses interact with cellulose microbrils and lignin to
form a unique natural composite structure [16]. Hosseinaei et al.
[17,18] investigated injection-molded PP composites lled with
hemicellulose-extracted Southern yellow pine by means of liquid
hot-water treatment (extraction at 140, 155, and 170 C for
60 min). They found decreased water absorption, increased tensile
properties, and improved mold resistance in the resulting compos-
ites. These improvements were attributed to the decreased hygro-
scopicity of the extracted wood. Pelaez-Samaniego et al. [19] used
hot water extracted ponderosa pine with bark to make HDPE and
PP composites. Water resistance of these composites was en-
hanced with unchanged or even improved mechanical properties
compared to the unextracted controls. Removal of hemicelluloses
by alkali treatment can improve the mechanical properties of lig-
nocellulosic bers [20,21] and composites including them
[2225]. This is because cellulose microbrils tend to come in
closer contact with one another, forming a more aggregated and
compact brillar structure when the hemicellulose that coats the
cellulose microbrils is removed [26].
http://dx.doi.org/10.1016/j.matdes.2014.02.018
0261-3069/ 2014 Elsevier Ltd. All rights reserved.

Corresponding author. Tel./fax: +86 451 82191993.

E-mail address: qwwang@nefu.edu.cn (Q. Wang).
Materials and Design 58 (2014) 339345
Contents lists available at ScienceDirect
Materials and Design
j our nal homepage: www. el sevi er . com/ l ocat e/ mat des
Cellulose ber is professed to be a better reinforcing ller than
wood ber. Owing to a complex network of inter- and intro-molec-
ular hydrogen bonds, cellulose chains have the ability to pack
themselves in a tight, high-strength and high stiffness crystalline
edice [27]. Ashori and Nourbakhsh [28] used microcrystalline cel-
lulose as a reinforcing agent in wood our/PP composites. Results
show that tensile, exural, and impact strengths of the composites
were signicantly enhanced with addition of microcrystalline cel-
lulose. Partial replacement of coarse spruce wood our with micro-
brillated cellulose (10 wt.%) resulted in improvements in almost
all mechanical properties of the resulting WPCs [29].
Our study was designed to determine the effects of the removal
of hemicellulose and lignin (together or respectively) on the
mechanical properties and water absorption of the resulting high
density polyethylene composites. The interfacial adhesion in com-
posites was characterized with scanning electron microscopy and
dynamical mechanical analysis.
2. Materials and methods
2.1. Materials
Wood our was prepared frompoplar (Populus ussuriensis Kom.)
sapwood chips in a hammer mill to pass through sieves measuring
80100 mesh. High density polyethylene (HDPE) pellets (5000S)
purchased from Daqing Petrochemical Co., China, bear a density
of 0.954 g cm
3
and a melt ow index of 0.7 g/10 min (190 C,
2.16 kg according to ASTM: D1238). The HDPE pellets were chilled
and ground to a ne powder for use. Maleated polyethylene
(MAPE, A-C

575A) was supplied by Honeywell International Inc.,

USA, with a MA grafting ratio of 3 wt.% and a saponication value
of 35 mg KOH/g.
2.2. Preparation of wood particles
Four types of wood particles were prepared from the
80100 mesh poplar our, following methods reported in the pre-
vious literature:
(1) WF Extracted wood particle: wood our was extracted in a
Soxhlet extractor with a mixture of ethanol and toluene (1:2
in volume) for 6 h to remove soluble extractives.
(2) HR Hemicellulose-removed wood particle: hemicelluloses
were removed from the WF according to TAPPI 203 [30],
leaving lignin and cellulose.
(3) HC Holocellulose (delignied wood particle): the WF was
delignied using a NaClO
2
treatment, leaving hemicelluloses
and cellulose [31].
(4) aC a-Cellulose (both hemicellulose and lignin removed
wood particle): the HC was further treated with a 17.5%
NaOH aqueous solution to remove hemicelluloses according
to TAPPI 203 [30].
These chemical treatments can potentially alter the size of
wood particles. To remove the confounding variable of particle
size, all particles were pulverized to 100160 mesh using an herb
grinder with knife. These particles were dried at 105 C for 24 h,
and then stored in a sealed container for use. In our previous paper,
the Fourier-transform infrared analysis of the wood particles
showed that hemicellulose was completely removed and a trace
residue of lignin was found in HC [32].
2.3. Preparation of composite specimens
Wood particles were dry-mixed with the HDPE powder and
MAPE at a specic ratio as shown in Table 1. They were com-
pounded using a co-rotating twin-screw extruder (diame-
ter = 8 mm and L/D = 40) equipped with a volumetric feeder and
a strand pelletizer (Leistritz ZSE-18, Leistritz Extrusionstechnik
GmbH, Germany). The extrusion temperature ranged from 150 to
175 C over eight zones along the extruder barrel. The extrudate
was cut into pellets using the strand pelletizer. The pellets were
then injection-molded (SE50D, Sumitomo Heavy Industries, Japan)
into standard specimens for mechanical testing. Injection and mold
temperatures were 180 and 50 C, respectively.
2.4. Characterization
2.4.1. Morphological analysis
The impact specimens were frozen in liquid nitrogen for 5 min
and broken by hand. The fractured surfaces were subsequently
dried, sputter-coated with gold, and then observed with a eld
emission scanning electron microscope (FE-SEM, Quanta 200F,
FEI Company, USA) at an accelerated voltage of 30 kV.
2.4.2. Mechanical test
Type-I dumbbell-shaped tensile specimens measuring
165 13 3 mm
3
were tested using a screw-driven universal
testing machine (Model 4466, Instron Inc., Canton, OH, USA) with
an 8.9 kN load cell according to ASTM: D638. The tests were con-
ducted at a crosshead speed of 5 mm min
1
with deections mea-
sured using a 50 mm extensometer (MTS 634.12E-24). Izod impact
strength was determined from specimens measuring
63.5 12.7 3 mm
3
using a BPI-0-1 Basic Pendulum Impact tester
(Dynisco, MA) according to ASTM: D256. All test specimens were
conditioned at 23 C and 65% relative humidity for 7 days prior
to testing. Six replicates were used for each formulation.
2.4.3. Water absorption test
Specimens measuring 30 20 3 mm
3
were cut from the ends
of dumbbell-shaped composites, oven-dried and then weighed.
The specimens were subsequently immersed in distilled water at
23 C for 2 months with a daily change of water. The weight and
thickness of each specimen were determined at stated periods.
Water absorption of each specimen was expressed as water weight
absorbed, divided by the dry weight of the specimen. The thickness
swelling of each specimen was calculated based on its dry dimen-
sions. Five replicates were used for each formulation.
2.4.4. Dynamic mechanical analysis (DMA)
Dynamic mechanical properties of the composites were
measured using a dynamic mechanical analyzer (DMA Q800, TA
Instruments, New Castle, USA). Tests were performed using
single-cantilever strain-controlled mode with an oscillating ampli-
tude of 15 lm and a frequency of 1 Hz. The temperature was swept
from 25 to 135 C at 3 C min
1
. The specimen dimension was
35 mm 12.0 mm 3 mm. Three replicates were used for each
formulation.
Table 1
Formulations of the composites.
Sample Fiber type Fiber (wt.%) HDPE (wt.%) MAPE (wt.%)
WF/PE WF 40 60 0
HR/PE HR 40 60 0
HC/PE HC 40 60 0
aC/PE aC 40 60 0
WF/PE/MA WF 40 58 2
HR/PE/MA HR 40 58 2
HC/PE/MA HC 40 58 2
aC/PE/MA aC 40 58 2
340 R. Ou et al. / Materials and Design 58 (2014) 339345
3. Results and discussion
3.1. Morphological analysis
Regardless of particle type, the fractured surfaces of the
composites without MAPE clearly showed boundaries and gaps be-
tween the HDPE matrix and the wood particles (Fig. 1a and b).
Numerous holes resulting from particle pullout were observed on
the fractured surfaces. The particles appeared to remain intact,
and were free from the adhering matrix after they were pulled
out. This result indicates insufcient interfacial bonding between
wood particles and HDPE. In contrast to the uncoupled composites,
composites with 2% MAPE had fewer gaps or voids on the fracture
surface (Fig. 1c and d). The particle/matrix interface became ob-
scure, and wood particles were fully covered by HDPE, as indicated
by the dashed circle in Fig. 1. Wood particles were fractured (ar-
rows in Fig. 1), indicating that the stress was effectively transferred
across the interface to the reinforcing particles. This may be have
been due to the interfacial interactions between the wood particles
and the MAPE [33]. Obvious differences in particle morphology and
the interface between the WF, HR, and HC-based composites could
not be discerned from the micrographs (micrographs for WF and
HC-based composites were not shown). The aC ber exhibited a
smaller diameter and greater aspect ratio than the other particles
(Fig. 1d). The cell structure was nearly indiscernible in the aC,
which also exhibited a much smoother surface than that of WF.
3.2. Mechanical properties
This study found that removal of wood cell wall composition af-
fects the tensile properties of composites (Fig. 2). Composites lled
with aC exhibited the highest tensile strength (r), elongation at
break (e), and toughness (energy required to break, EB). This may
be due to the fact that aC had the highest aspect ratio (7.54,
3.95, 3.67, and 3.32 for aC, HR, HC, and WF, respectively) and
degree of crystallinity (72.0%, 57.4%, 51.1%, and 43.5% for aC, HR,
HC, and WF, respectively). The presence of MAPE in the
formulations caused a signicant increase in the tensile strength,
tensile modulus (E), elongation at break, and toughness of the com-
posites (Table 2), especially the e and toughness. This may be due
to the formation of a soft interphase [34]. According to the classical
theory of mechanics, load applied to a particle-reinforced compos-
ite is transferred from the matrix to the particles by shear stress
along the particlematrix interface [35]. The incorporation of
MAPE promoted the dispersion of wood particles in the HDPE ma-
trix and improved the interfacial interaction between the wood
particles and the matrix, as shown by SEM analysis (Fig. 1). This re-
sults in more efcient stress transfer from the matrix to the wood
particles, leading to a signicant improvement in the mechanical
properties of the composites. This indicates that good interfacial
adhesion is the key to obtaining signicant improvement in
mechanical properties.
Compared to WF/PE, removal of hemicellulose caused 9.2%,
23.4%, and 33.2% higher r, e, and EB, respectively (Table 2). This
may be explained by the fact that removal of hemicellulose caused
Fig. 1. SEM micrographs of fracture surface of the composites (a) HR/PE, (b) aC/PE, (c) HR/PE/MA, (d) aC/PE/MA.
0 1 2 3 4 5 6 7
0
8
16
24
32
40
48
C/PE
HR/PE
WF/PE
HC/PE
C/PE/MA
HR/PE/MA
WF/PE/MA
HC/PE/MA
S
t
r
e
s
s

(
M
P
a
)
Strain (%)
Fig. 2. Stressstrain curves of the composites.
R. Ou et al. / Materials and Design 58 (2014) 339345 341
a reduced polarity of HR. This in turn may promote the dispersion
of HR in the HDPE and improve the interfacial bonding between HR
and HDPE, which increases elongation at break. However, the low-
er tensile modulus for HR/PE was unexpected, because the crystal-
linity of HR is higher than that of WF (not shown). Further
investigation is necessary to elucidate this result. When the HR/PE
composite was coupled with MAPE, there was a slight increase in r
and E. However, there was a reduction of 11.3% and 7.4% in e and
EB, respectively than the MAPE coupled WF/PE composite. This re-
sult is opposite to that of the uncoupled composites. The increased
r and E may be due to the higher crystallinity of HR than that of
WF. Elongation at break depends on the strain of the wood particle
and also the matrix. The higher crystallinity of HR makes the com-
posites more rigid and experiencing decreased strain.
Results also show that delignication did not cause signicant
changes in r, e, and EB, while E increased by 13% compared to
WF based composites. The increase in E may be explained by the
fact that when the spherical lignin in the compound middle lamel-
la and lens-shaped lignin in secondary wall were removed, the
porosity in cell walls increased [36]. The HC particles were com-
promised by the combined inuence of high temperature, com-
pression and shear stress during extrusion/injection processing.
Therefore, the HC cell walls exhibited greater deformation than
the other three particles. This may also be proved by the highest
density of HC-based composites (Table 2). Consequently, the highly
compressed HC endowed the resulting composites with a high
modulus. In the presence of MAPE, HC/PE/MA showed comparable
behavior, with a 12.0% and 14.6% higher e and EB than WF/PE/MA,
respectively.
Compared to WF/PE, aC/PE had 13.4%, 50.4%, and 70.7% higher
r, e, EB, respectively (Table 2). In the presence of MAPE, aC/PE/
MA had 19.7%, 18.7%, and 41.5% higher r, e, EB than WF/PE/MA,
respectively. This may be due to the high aspect ratio of the aC,
which is positively related with stress transfer efciency [35]. It
is expected that removal of both hemicellulose and lignin would
increase crystallinity and create more rigid composites [37]. How-
ever, this was not the case. In fact, aC/PE exhibited an 8% lower E
compared to WF/PE, which is possibly due to the lower degree of
crystallinity (not shown) and density (Table 2) of aC/PE.
There is a remarkably different response in notched impact (NI)
and unnotched impact strengths (UNI) of the uncoupled and MAPE
coupled composites (Table 2). For uncoupled composites, the UNI
strength is approximately double the NI strength, while the UNI
strength is 56 times of the NI strength for the coupled composites.
Furthermore, the difference in UNI strength between uncoupled
and coupled composites is much larger than that in NI strength.
The UNI strength of a composite consists of the crack initiation
and crack propagation energy, and the former dominates [38],
whereas NI strength is mainly attributed to the latter [39].
The energy difference between UNI and NI strength is the crack
initiation energy. The crack propagation energy is the accumulative
work of various mechanisms that dissipate energy, including the
work-of-fracture of the reinforcing particle and the matrix and
the energy dissipated from different particlematrix interactions
(sliding, debonding, particle pullout, etc.) [40]. For uncoupled com-
posites, the voids or gaps between the wood particle and matrix act
as stress concentration sites for crack initiation, decreasing the
crack initiation energy signicantly [41]. Accordingly, the differ-
ence between UNI and NI strength is small. With the incorporation
of MAPE, the crack initiation energy increased signicantly, due to
improved wood particlematrix interfacial adhesion [42], thereby
increasing UNI strength.
The HC/PE exhibited the lowest UNI and NI strength compared
to the other uncoupled composites. This may be explained by the
fact that delignication can reduce its strength [43] as well as
the wettability from the HDPE which results from increased polar-
ity of HC. Furthermore, the increased polarity created an increase
in possible particle agglomeration, in turn creating more regions
of stress concentration that require less energy to initiate a crack.
Therefore, the crack initiation energy, work-of-fracture of HC,
and the energy dissipated from HCHDPE interactions resulting
from particle pullout decreased. Removal of hemicellulose can
increase its hydrophilicity and the strength of HR [21]. This indi-
cates an increase in the crack initiation energy, work-of-fracture
of HR, and the energy dissipated from HRHDPE interactions, all
of which are responsible for higher UNI and NI strength of HR/PE
as compared to WF/PE. Our study found that aC/PE had the highest
UNI and NI strength, which may be attributed to the higher aspect
ratio and strength of aC. Pullout of the long bers dissipates high
energy, and strong bers impart high work-of-fracture. There
was no signicant difference in both UNI and NI strength between
the composites lled with WF, HR, and HC in the presence of MAPE.
3.3. Water absorption
A layer of HDPE in the injecting-molded specimen can inhibit
the water absorption of WPCs [44]. The rate of water absorption
was high in the initial stage, and then gradually slowed down
(Fig. 3). The nal water absorption and thickness swelling of com-
posites with and without MAPE decreased in the following order:
HC > aC > WF > HR (Table 2). The water absorption and thickness
swelling of the coupled composites were considerably reduced
by adding MAPE, and the difference between the composites
decreased (Fig. 3). The decrease in water absorption and thickness
swelling may result from the improved dispersion of wood
particles in the HDPE matrix as well as the increased interfacial
adhesion between the wood particles and the matrix caused by
MAPE. The improved interfacial adhesion can decrease the water
diffusion rate between the wood particles [45]. The reduced
amount of hygroscopic hydroxyl groups in the wood polymers
due to esterication with anhydride groups of MAPE should not
reduce water absorption. This is because an equal amount of
Table 2
Mechanical and physical properties of the composites: r = tensile strength, E = tensile modulus, e = elongation at break, EB = energy to break, UNI = unnotched impact strength,
NI = notched impact strength, WA = water absorption at 61 days, TS = thickness swell at 61 days.
Samples r (MPa) E (GPa) e (%) EB (J) UNI (kJ m
2
) NI (kJ m
2
) WA (%) TS (%) Density (g cm
3
)
WF/PE 23.8 (0.2) 4.00 (0.12) 2.18 (0.19) 43.0 (4.8) 5.92 (0.53) 2.98 (0.08) 6.26 (0.13) 3.77 (0.15) 1.091 (0.00)
HR/PE 26.0 (0.4) 3.57 (0.08) 2.69 (0.31) 57.3 (5.3) 7.28 (0.35) 3.32 (0.04) 3.37 (0.16) 2.30 (0.13) 1.092 (0.00)
HC/PE 24.6 (0.2) 4.52 (0.10) 2.21 (0.21) 48.7 (2.2) 5.48 (0.25) 2.41 (0.06) 8.78 (0.22) 6.18 (0.18) 1.095 (0.00)
aC/PE 27.0 (0.3) 3.68 (0.08) 3.23 (0.27) 73.4 (6.6) 11.80 (0.62) 6.29 (0.36) 7.84 (0.13) 4.88 (0.13) 1.086 (0.00)
WF/PE/MA 40.1 (0.3) 4.19 (0.08) 6.10 (0.44) 215.6 (6.7) 27.46 (2.40) 5.44 (0.14) 2.72 (0.04) 1.03 (0.05) 1.092 (0.00)
HR/PE/MA 42.1 (0.2) 4.28 (0.07) 5.41 (0.74) 199.6 (29.1) 26.46 (0.67) 5.49 (0.20) 2.13 (0.05) 0.99 (0.02) 1.093 (0.00)
HC/PE/MA 40.6 (0.5) 4.94 (0.17) 6.83 (0.34) 247.1 (15.9) 28.25 (0.87) 5.41 (0.2) 3.00 (0.07) 1.82 (0.03) 1.100 (0.00)
aC/PE/MA 48.0 (0.6) 4.37 (0.11) 7.24 (0.71) 305.1 (19.7) 38.85 (1.25) 6.67 (0.48) 2.76 (0.12) 1.23 (0.08) 1.099 (0.00)
Numbers in parenthesis denote standard deviations.
342 R. Ou et al. / Materials and Design 58 (2014) 339345
hygroscopic carboxyl group (COOH) is formed [46], and the num-
ber of anhydride groups is much lower than the total number of
hydroxyl groups in the composites.
The hygroscopicity of wood particles, which is highly correlated
with the free OH groups of their three main compositions, is
mainly responsible for water absorption in WPCs. Hemicelluloses
had the highest capacity to absorb water, followed by cellulose
and then lignin. Removal of hemicellulose decreased the accessible
hydroxyl groups of HR. Therefore, HR should be properly wetted by
HDPE; water diffusion in the composites can therefore be slowed
down (Fig. 3) [17,19].
It was observed that delignication of wood particle
signicantly increased water absorption and thickness swelling
in both uncoupled and coupled composites (Fig. 3). This was to
be expected, since lignin is hydrophobic and HC is more hydro-
philic and porous than wood particles containing lignin [36].
Swelling of the composites during water immersion was the result
of swelling of wood cell walls. Compared to WF/PE, the increase in
nal thickness swelling (63.9%) was much higher than the increase
in water absorption (40.2%) for HC/PE (Table 2). One explanation
could be that the HC was highly compressed during
extrusion/injection processing, as noted above. Consequently, the
compressed HC will spring back after absorbing water, thereby cre-
ating more thickness swelling. HC/PE/MA exhibited the same
behavior as the HC/PE (Table 2). When both lignin and hemicellu-
lose were removed, more free hydroxyl groups were exposed on
the surface of cellulose crystallites [47]. This resulted in higher
water absorption and thickness swelling compared to WF based
composites with and without MAPE (Fig. 3).
The higher aspect ratio of aC (7.54) could be another
explanation for the higher water absorption and thickness
swelling. The aspect ratio of WF is 3.32. Previous studies show that
the high aspect ratio of ber increases water absorption and
thickness swelling [44]. Shorter bers may be more easily coated
by the matrix. Use of longer bers may create more voids within
the composites, thereby increasing water content [44].
3.4. Dynamic mechanical analysis
A decreasing trend in the storage modulus over the entire
temperature range was observed (Fig. 4). This may have been
due to the increased mobility of chain segments of HDPE at high
temperatures. Incorporation of stiff wood particles into the HDPE
matrix signicantly increased the storage modulus (E
0
) of the
resulting composites compared to the neat HDPE. This results from
the stiffness of wood particle [48]. Coupled composites had a high-
er E
0
than uncoupled composites at low temperatures. This greater
stiffness after the addition of MAPE resulted from the improved
interfacial adhesion, since stress was more efciently transferred
from the matrix to wood particles [49], as shown in the tensile
modulus (Table 2). The stiffness of composites at room tempera-
ture (above T
g
) depends on the density of the composites and the
stiffness of both the matrix (crystallinity) and the particles. The
stiffness of particles will dominate at high temperatures [50].
At room temperature for uncoupled composites, HC/PE had the
highest E
0
, followed by WF/PE, aC/PE, and HR/PE, which accords
with the results for the tensile modulus. The higher E
0
of HC/PE is
associated with its higher density and the higher level of deforma-
tion of HC, as noted above. aC had the highest crystallinity among
the four wood particles (not shown), but did not contribute to the
highest E
0
for aC/PE. This likely resulted from the lowest crystallin-
ity (not shown) and density (Table 2) of aC/PE. When the temper-
ature was close to the melting point of HDPE, results differed
signicantly from those at room temperature, since the composites
exhibited a decreasing order of E
0
as: aC/PE > HR/PE > WF/PE > HC/
PE. For coupled composites, E
0
decreased in the following order:
HC/PE/MA > aC/PE/MA > HR/PE/MA > WF/PE/MA at room tempera-
ture, which is consistent with the results for the tensile modulus
0
2
4
6
8
HC/PE C/PE
WF/PE HR/PE
C/PE/MA
HR/PE/MA
W
a
t
e
r

a
b
s
o
r
p
t
i
o
n

(
%
)
HC/PE/MA
WF/PE/MA
(a)
0 10 20 30 40 50 60
0
1
2
3
4
5
6
S
w
e
l
l
i
n
g

(
%
)
Time (d)
(b)
Fig. 3. (a) Water absorption (b) and thickness swelling of the composites.
3000
2000
1000
0
4000
40 60 80 100 120
Temperature (C)
0.3
0.2
0.1
T
a
n

0.3
0.2
0.1
T
a
n

(
M
P
a
)
3000
2000
1000
0
4000
E

(
M
P
a
)
HDPE
WF/PE
HR/PE
HC/PE
C/PE
HDPE
WF/PE/MA
HR/PE/MA
HC/PE/MA
C/PE/MA
(a)
(b)
Fig. 4. Storage modulus (E
0
) and mechanical damping (tand) of the composites as a
function of temperature: (a) without MAPE and (b) with MAPE.
R. Ou et al. / Materials and Design 58 (2014) 339345 343
(Table 2). The lowest E
0
of WF/PE/MA may result from its lowest
crystallinity (not shown). The lower E
0
of HC based composites at
higher temperature likely resulted from HCs higher exibility
compared to other particle types.
The damping factor tand (E
0 0
/E
0
) amplitude among all the com-
posites was almost the same at room temperature (Fig. 4). The dif-
ference increased with an increase in temperature. The uncoupled
composites exhibited a decreasing order of tand amplitude as: HC/
PE > WF/PE > HR/PE > aC/PE at higher temperatures. The high stiff-
ness of the particle restrains the segment mobility of the matrix
molecules, resulting in a low tand amplitude of the composites
[51]. This indicates that the wood particles exhibited a decreasing
order of exibility as: HC > WF > HR > aC. The composites with
MAPE showed a similar trend (Fig. 4b). In the temperature range
50 C to 120 C, the tand amplitude of the composites with MAPE
was slightly higher than that of the composites without MAPE.
A composite with poorer interfacial adhesion between particles
and matrix tends to dissipate more energy, showing a higher tand
magnitude than a composite with a strongly bonded interface [52].
On the other hand, the low molecular weight of MAPE can increase
the mobility of molecules and in turn increase the tand magnitude
[53]. The latter was dominant at temperatures between 50 C and
120 C, resulting in higher tand magnitude for the coupled com-
posites. At temperatures over 120 C, the chain mobility of the bulk
matrix in the uncoupled composites increased, while the molecular
motion in the interfacial region was restricted in the coupled com-
posites due to strong adhesion. Accordingly, the former began to
dominate, resulting in a higher tan d magnitude for uncoupled
composites at higher temperatures.
4. Conclusions
Results of this study can be summarized with the following key
conclusions:
(1) Removal of hemicellulose can considerably reduce water
absorption and thickness swelling, thereby dimensionally
stabilizing the WPC. Tensile strength, elongation at break,
toughness, and unnotched and notched impact strength also
improved in composites without MAPE.
(2) Removal of lignin caused the highest tensile modulus and
storage modulus for the resulting composites at room tem-
perature. But the lowest water resistance, thermal stability,
and storage modulus at high temperatures was observed
after lignin removal.
(3) Removal of both hemicellulose and lignin resulted in the
highest tensile strength, elongation at break, toughness,
and impact strength of the composites, regardless of the
incorporation of MAPE.
Acknowledgments
The supports from the National Key Technologies R&D Program
of China (No. 2012BAD32B04) and National Natural Science Foun-
dation of China (Nos. 31010103905 and 31070507) are gratefully
acknowledged. Yanjun Xie thanks the Program for New Century
Excellent Talents in University of Ministry of Education of China
(No. NCET-11-0608).
References
[1] Wambua P, Ivens J, Verpoest I. Natural bres: can they replace glass in bre
reinforced plastics? Compos Sci Technol 2003;63(9):125964.
[2] Alsaeed T, Yousif BF, Ku H. The potential of using date palm bres as
reinforcement for polymeric composites. Mater Des 2013;43:17784.
[3] Saheb DN, Jog JP. Natural ber polymer composites: a review. Adv Polym
Technol 1999;18(4):35163.
[4] Ashori A. Wood-plastic composites as promising green-composites for
automotive industries! Bioresour Technol 2008;99(11):46617.
[5] Canetti M, Bertini F. Supermolecular structure and thermal properties of poly
(ethylene terephthalate)/lignin composites. Compos Sci Technol 2007;67:
31517.
[6] Morandim-Giannetti AA, Agnelli JAM, Lanas BZ, Magnabosco R, Casarin SA,
Bettini SH. Lignin as additive in polypropylene/coir composites: thermal,
mechanical and morphological properties. Carbohyd Polym 2012;87:25638.
[7] Pouteau C, Dole P, Cathala B, Averous L, Boquillon N. Antioxidant properties of
lignin in polypropylene. Polym Degrad Stabil 2003;81:918.
[8] Wood BM, Coles SR, Maggs S, Meredith J, Kirwan K. Use of lignin as a
compatibiliser in hemp/epoxy composites. Compos Sci Technol
2011;71:180410.
[9] Rozman H, Tan K, Kumar R, Abubakar A, Mohd Ishak Z, Ismail H. The effect of
lignin as a compatibilizer on the physical properties of coconut ber-
polypropylene composites. Eur Polym J 2000;36:148394.
[10] Guigo N, Vincent L, Mija A, Naegele H, Sbirrazzuoli N. Innovative green
nanocomposites based on silicate clays/lignin/natural bres. Compos Sci
Technol 2009;69:197984.
[11] Stewart D. Lignin as a base material for materials applications: chemistry,
application and economics. Ind Crop Prod 2008;27:2027.
[12] Karimi A, Nazari S, Ghasemi I, Tajvidi M, Ebrahimi G. Effect of the
delignication of wood bers on the mechanical properties of wood ber
polypropylene composites. J Appl Polym Sci 2006;102:475963.
[13] Fabiyi JS, McDonald AG, McIlroy D. Wood modication effects on weathering
of HDPE-based wood plastic composites. J Polym Environ 2009;17:3448.
[14] Fabiyi JS, McDonald AG. Weathering performance of delignied pine-based
polyvinyl chloride composites. J Reinf Plast Comp 2013;32:54763.
[15] Pickering KL, Beg MDH. The effects of residual lignin content on wood bre
reinforced polypropylene composites. Adv Mater Res 2007;29:3236.
[16] Scheller HV, Ulvskov P. Hemicelluloses. Plant Biology 2010;61:26389.
[17] Hosseinaei O, Wang S, Enayati AA, Rials TG. Effects of hemicellulose extraction
on properties of wood our and woodplastic composites. Compos Part A
Appl S 2012;43:68694.
[18] Hosseinaei O, Wang S, Taylor AM, Kim J-W. Effect of hemicellulose extraction
on water absorption and mold susceptibility of wood-plastic composites. Int
Biodeter Biodegr 2012;71:2935.
[19] Pelaez-Samaniego Manuel R, Yadama V, Lowell E, Amidon Thomas E, Chaffee
Timothy L. Hot water extracted wood ber for production of wood plastic
composites (WPCs). Holzforschung 2013;67:193200.
[20] Mwaikambo LY, Ansell MP. Modication of hemp, sisal, jute, and kapok bres
by alkalization. J Appl Polym Sci 2002;84:222234.
[21] Roy A, Chakraborty S, Kundu SP, Basak RK, Basu Majumder S, Adhikari B.
Improvement in mechanical properties of jute bres through mild alkali
treatment as demonstrated by utilization of the Weibull distribution model.
Bioresour Technol 2012;107:2228.
[22] Venkateshwaran N, ElayaPerumal A, Arunsundaranayagam D. Fiber surface
treatment and its effect on mechanical and visco-elastic behaviour of banana/
epoxy composite. Mater Des 2014;47:1519.
[23] Yousif BF, Shalwan A, Chin CW, Ming KC. Flexural properties of treated and
untreated kenaf/epoxy composites. Mater Des 2012;40:37885.
[24] Ray D, Sarkar BK, Bose NR. Impact fatigue behaviour of vinylester resin matrix
composites reinforced with alkali treated jute bres. Compos Part A Appl S
2002;33:23341.
[25] Ray D, Sarkar BK, Rana A, Bose NR. The mechanical properties of vinylester
resin matrix composites reinforced with alkali-treated jute bres. Compos Part
A-Appl S 2001;32:11927.
[26] Wan JQ, Wang Y, Xiao Q. Effects of hemicellulose removal on cellulose ber
structure and recycling characteristics of eucalyptus pulp. Bioresour Technol
2010;101:457783.
[27] Jarvis M. Chemistry: cellulose stacks up. Nature 2003;426:6112.
[28] Ashori A, Nourbakhsh A. Performance properties of microcrystalline cellulose
as a reinforcing agent in wood plastic composites. Compos Part B Eng
2010;41:57881.
[29] Turku I, Krki T. Reinforcing woodplastic composites with macro-and micro-
sized cellulosic llers: comparative analysis. J Reinf Plast Compos
2013;32:174656.
[30] TAPPI Test Method T 203 om-93. Alpha-, beta- and gamma-cellulose in pulp.
Atlanta, Georgia, USA: TAPPI Press; 1994.
[31] Wise LE, Murphy M, DAddieco AA. Chlorite holocellulose, its fractionation and
bearing on summative wood analysis and on studies on the hemicelluloses.
Pap Trade J 1964;122:3543.
[32] Ou R, Xie Y, Wolcott MP, Yuan F, Wang Q. Effect of wood cell wall composition
on the rheological properties of wood particle/high density polyethylene
composites. Compos Sci Technol 2014;93:6875.
[33] Paunikallio T, Kasanen J, Suvanto M, Pakkanen TT. Inuence of maleated
polypropylene on mechanical properties of composite made of viscose ber
and polypropylene. J Appl Polym Sci 2003;87:1895900.
[34] Oksman K, Lindberg H. Inuence of thermoplastic elastomers on adhesion in
polyethylene-wood our composites. J Appl Polym Sci 1998;68:184555.
344 R. Ou et al. / Materials and Design 58 (2014) 339345
[35] Gibson RF. Principles of composites material mechanics. New York: McGraw-
Hill; 1994.
[36] Hafrn J, Fujino T, Itoh T, Westermark U, Terashima N. Ultrastructural changes
in the compound middle lamella of Pinus thunbergii during lignication and
lignin removal. Holzforschung 2000;54:23440.
[37] Rowell RM, Sanadi AR, Cauleld DF, Jacobson RE. Utilization of natural bers in
plastic composites: problems and opportunities. In: Lea AL, Carvalho FX,
Frollini E, editors. Lignocellulosicplastic composites. Sao Paulo: Universdade
de Sao Paulo; 1997. p. 2352.
[38] Rajabian M, Dubois C. Polymerization compounding of HDPE/Kevlar
composites. I. Morphology and mechanical properties. Polym Compos
2006;27:12937.
[39] Yang H-S, Wolcott MP, Kim H-S, Kim S, Kim H-J. Effect of different
compatibilizing agents on the mechanical properties of lignocellulosic
material lled polyethylene bio-composites. Compos Struct 2007;79:36975.
[40] Lpez JP, Girons J, Mendez JA, Plach MA, Vilaseca F, Mutj P. Impact and
exural properties of stone-ground wood pulp-reinforced polypropylene
composites. Polym Compos 2013;34:8428.
[41] Ou R, Zhao H, Sui S, Song Y, Wang Q. Reinforcing effects of Kevlar ber on the
mechanical properties of wood-our/high-density-polyethylene composites.
Compos Part A Appl S 2010;41:12728.
[42] Stark N, Rowlands R. Effects of wood ber characteristics on mechanical
properties of wood/polypropylene composites. Wood Fiber Sci
2003;35:16774.
[43] Mukherjee A, Ganguly PK, Sur D. Structural mechanics of jute: the effects of
hemicellulose or lignin removal. J Text 1993;I(84):34853.
[44] Migneault S, Koubaa A, Erchiqui F, Chaala A, Englund K, Wolcott MP. Effects of
processing method and ber size on the structure and properties of wood
plastic composites. Compos Part A Appl S 2009;40:805.
[45] Xie Y, Xiao Z, Grneberg T, Militz H, Hill CAS, Steuernagel L, et al. Effects of
chemical modication of wood particles with glutaraldehyde and 1,3-
dimethylol-4,5-dihydroxyethyleneurea on properties of the resulting
polypropylene composites. Compos Sci Technol 2010;70:200311.
[46] Gao H, Xie Y, Ou R, Wang Q. Grafting effects of polypropylene/polyethylene
blends with maleic anhydride on the properties of the resulting woodplastic
composites. Compos Part A Appl S 2012;43:1507.
[47] Ichazo MN, Albano C, Gonzlez J, Perera R, Candal MV. Polypropylene/wood
our composites: treatments and properties. Compos Struct 2001;54:20714.
[48] Mohanty S, Verma SK, Nayak SK. Dynamic mechanical and thermal properties
of MAPE treated jute/HDPE composites. Compos Sci Technol 2006;66:53847.
[49] Hristov V, Vasileva S. Dynamic mechanical and thermal properties of modied
poly(propylene) wood ber composites. Macromol Mater Eng
2003;288:798806.
[50] Manikandan Nair KC, Thomas S, Groeninckx G. Thermal and dynamic
mechanical analysis of polystyrene composites reinforced with short sisal
bres. Compos Sci Technol 2001;61:251929.
[51] Lopez Manchado MA, Biagiotti J, Kenny JM. Comparative study of the effects of
different bres on the processing and properties of ternary composites based
on PP-EPDM blends. Polym Compos 2002;23:77989.
[52] Mohanty S, Nayak SK. Interfacial, dynamic mechanical, and thermal ber
reinforced behavior of mape treated sisal ber reinforced HDPE composites. J
Appl Polym Sci 2006;102:330615.
[53] Nyambo C, Mohanty AK, Misra M. Polylactide-based renewable green
composites from agricultural residues and their hybrids. Biomacromolecules
2010;11:165460.
R. Ou et al. / Materials and Design 58 (2014) 339345 345